Software-guided insulin dosing improves intrapartum glycemic management in women with diabetes mellitus

During labor, maintenance of maternal euglycemia is critical to decrease the risk of neonatal hypoglycemia and associated morbidities. When continuous intravenous insulin infusion is needed, standardized insulin dosing charts have been used for titration of insulin to maintain glucose in target range. The GlucoStabilizer software program (Indiana University Health Inc, Indianapolis, IN) is a software-guided insulin dosing system that calculates the dose of intravenous insulin that is needed based on metabolic parameters, target glucose concentration, and an individual's response to insulin. Although this tool has been validated and shown to reduce both hypoglycemia and errors in critical care settings, the utility of this software has not been examined in obstetrics.The purpose of this study was to determine whether the use of intravenous insulin dosing software in women with pregestational or gestational diabetes mellitus that requires intrapartum insulin infusion can improve the rate of glucose concentration in target range (70-100 mg/dL; 3.9-5.5 mmol/L) at the time delivery.We performed a retrospective cohort study comparing laboring patients with diabetes mellitus that required insulin infusion who were dosed by standard insulin dosing chart vs the GlucoStabilizer software program from January 2012 to December 2017. The GlucoStabilizer software program, which was implemented in May 2016, replaced the standard intravenous insulin dosing chart. Inclusion criteria were women with pregestational or gestational diabetes mellitus who were treated with an intravenous insulin infusion intrapartum for at least 2 hours. Maternal characteristics, glucose values in labor, and neonatal outcomes were extracted from delivery and neonatal records. The primary outcome was the percentage of women who achieved the target glucose range (defined as a blood glucose between 70-100 mg/dL; 3.9-5.5 mmol/L) before delivery. Parametric and nonparametric statistics were used to compare both groups; a probability value of.05 was considered statistically significant.We identified 22 patients who were dosed by a standard insulin dosing chart and 11 patients who were dosed by the GlucoStabilizer software program during intrapartum management. The GlucoStabilizer software program was superior in achieving glucose values in target range at delivery (81.8% vs 9.1%; P.001) compared with standard insulin dosing without increasing maternal hypoglycemia (0% vs 4.3%; P=.99). Patients whose insulin dosing was managed by the GlucoStabilizer software program also had lower mean capillary blood glucose values compared with the standard insulin infusion (102.9±5.9 mg/dL [5.7±0.33 mmol/L] vs 121.7±5.9 mg/dL [6.8±0.33 mmol/L]; P=.02). Before the initiation of the infusion, both groups demonstrated mean capillary blood glucose values outside of target range (122.6±8.8 mg/dL [6.7±0.49 mmol/L] for the GlucoStabilizer software program vs 131.9±10.1 mg/dL [7.3±0.56 mmol/L] for standard insulin treatment group; P=not significant). There were no significant differences in baseline maternal characteristics between the groups or neonatal outcomes.This study is the first to demonstrate that the use of software-guided intravenous insulin dosing in obstetrics can improve intrapartum glycemic management without increasing hypoglycemia in women with both pregestational and gestational diabetes mellitus that is treated with an insulin infusion.